I have never investigated baffles that are the same width as the voice coil of a tweeter or the diameter of the cone om a midrange driver so I do not know the answer to your question.
Theoretically the driver would be omni directional, but that is not really the case since the tweeter only plays in one direction if its a dome tweeter.
PS Audio's FR30 is absolutely effected since their planar mid and tweeter by nature only plays in halfspace.
So do the lower mid too until you hit the point where it starts to play more and more in full space until it fully plays 360° below 100 Hz.
Basically all speakers that goes from half space spread pattern to full space spread pattern are plagued by ths baffle step and need some compensation.
Only open baffle or dipole speakers that plays fullrange both forward and backwards are more or less free from baffle step effect and need no compensation.
Magnepan, Martin logan or Linkwitz dipole are good examples.
By the way, the loudspeaker measured in last post are B&O Penta which actually plays very frequency linear in the room down to 30Hz.
Can't play loud because of small drivers, but it is surprizingly well balanced and stretches deep in the bass. Baffle width are 140 mm.
Using them for my TV setup.
Theoretically the driver would be omni directional, but that is not really the case since the tweeter only plays in one direction if its a dome tweeter.
PS Audio's FR30 is absolutely effected since their planar mid and tweeter by nature only plays in halfspace.
So do the lower mid too until you hit the point where it starts to play more and more in full space until it fully plays 360° below 100 Hz.
Basically all speakers that goes from half space spread pattern to full space spread pattern are plagued by ths baffle step and need some compensation.
Only open baffle or dipole speakers that plays fullrange both forward and backwards are more or less free from baffle step effect and need no compensation.
Magnepan, Martin logan or Linkwitz dipole are good examples.
By the way, the loudspeaker measured in last post are B&O Penta which actually plays very frequency linear in the room down to 30Hz.
Can't play loud because of small drivers, but it is surprizingly well balanced and stretches deep in the bass. Baffle width are 140 mm.
Using them for my TV setup.
Last edited:
Could you explain this?
I would expect -6 dB with each doubling of distance (that is: not including baffle, diaphragm dimension and room effects).
With room effect in the lower bass region the reduction is less, maybe -3 to -4 dB, in the range you mentioned.
And for the transition from 2pi to 4pi radiation, below baffle step frequency the effect is more than -6 dB, but only for near distances.
-3 dB would be expected for a line source, right?
Maybe I just did not understand your explanation - thanks in advance!
Sorry for off topic, marcop999!
@stv
Most think the way you do so it's common misunderstanding.
We are measuring the acoustic power and acoustic power decrease with 3dB for every doubling of distance and increase with 3dB by every halfing of distance.
https://pulsarinstruments.com/news/understanding-decibels-decibel-scale-and-noise-measurement-units/
Theoretically the transition from 2pi to 4pi is always -6dB exactly.
In practice it's around -4dB because of reflections from walls, ceiling and floor.
When measuring at different distances it's obvious that amplitude level decreases with 3dB for every doubling of distance.
Someone more educated than me for sure can describe it better.
I'm just telling what I see and have learnt by practice.
Most think the way you do so it's common misunderstanding.
We are measuring the acoustic power and acoustic power decrease with 3dB for every doubling of distance and increase with 3dB by every halfing of distance.
https://pulsarinstruments.com/news/understanding-decibels-decibel-scale-and-noise-measurement-units/
Theoretically the transition from 2pi to 4pi is always -6dB exactly.
In practice it's around -4dB because of reflections from walls, ceiling and floor.
When measuring at different distances it's obvious that amplitude level decreases with 3dB for every doubling of distance.
Someone more educated than me for sure can describe it better.
I'm just telling what I see and have learnt by practice.
We are usually not measuring acoustic power but sound pressure.
(Total) Acoustic power is not reduced at all with distance, if I'm not mistaken.
(Edit: power is actually reduced by air absorption losses)
Edit2: I'll probably think over it an do some research to avoid misconception!
Anyway thanks for the response!
I don't think you are right about that statement.
Sorry, but don't tell me that I'm wrong when I for a fact know that measured soundlevel with a microphone and a software like REW or LMS that I use clearly reduce dB level with 3 dB for every doubling of distance.
Sorry, but don't tell me that I'm wrong when I for a fact know that measured soundlevel with a microphone and a software like REW or LMS that I use clearly reduce dB level with 3 dB for every doubling of distance.
Apologies for my wording, I did not want to be snarky.
I'll have to do some further research.
Thank you @flex2,
Let me do some consideration about your measures:
1) your room seem well treated and low susceptibility to external resonances caused by wooden furnishings and this simplify the measures.
2) The B&O Penta compared to other loadspeakers seem less affect to baffle step issue.
However now suppose we want compensate the baffle step, the question is at what distance you must use at reference ?
Suppose you compensate the baffle step at 50 cm adjusting the bass inductor. What happen to other distance ?
I don't think you can compensate equally at all distances.
So I guess if you compensate it at 50cm distance and repeat all the measurements , you will find that if before at 120cm the response was relatively flat, after compensation it will not be. Do you confirm ?
So if I wanted to make a more practical point, with your B&O Penta loudspeakers I wouldn't make any attempt to compensate for the baffle by modifying the crossover because I assume the listening point is more than a meter and I don't think 50 cm. Do you confirm ?
Apart this I have another question at what height did you put the microphone ?
I ask this becouse I see that frequency response of tweeter is relatively flat at 50cm but starting at 110cm starts to get lower.
I suppose this depend by off-axis dispersion tweeter. It is true ?
While the baffle step occurs at small distances therefore rather far from the listening point, instead the tweeter lowering occurs near the listening point.
So it seems that in the graph you reported the off-axis dispersion should have perhaps more attention then the baffle step compensation.
What you think about ?
Let me do some consideration about your measures:
1) your room seem well treated and low susceptibility to external resonances caused by wooden furnishings and this simplify the measures.
2) The B&O Penta compared to other loadspeakers seem less affect to baffle step issue.
However now suppose we want compensate the baffle step, the question is at what distance you must use at reference ?
Suppose you compensate the baffle step at 50 cm adjusting the bass inductor. What happen to other distance ?
I don't think you can compensate equally at all distances.
So I guess if you compensate it at 50cm distance and repeat all the measurements , you will find that if before at 120cm the response was relatively flat, after compensation it will not be. Do you confirm ?
So if I wanted to make a more practical point, with your B&O Penta loudspeakers I wouldn't make any attempt to compensate for the baffle by modifying the crossover because I assume the listening point is more than a meter and I don't think 50 cm. Do you confirm ?
Apart this I have another question at what height did you put the microphone ?
I ask this becouse I see that frequency response of tweeter is relatively flat at 50cm but starting at 110cm starts to get lower.
I suppose this depend by off-axis dispersion tweeter. It is true ?
While the baffle step occurs at small distances therefore rather far from the listening point, instead the tweeter lowering occurs near the listening point.
So it seems that in the graph you reported the off-axis dispersion should have perhaps more attention then the baffle step compensation.
What you think about ?
Last edited:
@stv
I can check if gated measurement behaves differently. It shouldn't, since the 3dB decreas is valid for measurement from 5cm to 10 cm too and then room have very little impact.
You are correct that it is measered with continous sinewave since I wanted to see result down to 20 Hz.
Use gated measurement for near field measurement above 200 Hz but have never reflected over any differences in result because of type of measurement.
But I check and let you know about what I find.
I can check if gated measurement behaves differently. It shouldn't, since the 3dB decreas is valid for measurement from 5cm to 10 cm too and then room have very little impact.
You are correct that it is measered with continous sinewave since I wanted to see result down to 20 Hz.
Use gated measurement for near field measurement above 200 Hz but have never reflected over any differences in result because of type of measurement.
But I check and let you know about what I find.
At such close distance the diaphragm behaves like a "big" surface radiator. once the distance gets big compared to the diaphragm diameter (maybe 50 cm to 1 m) it is more like a point source with 6 dB anechoic decrease with double distance.
I practice it like this...
I know that the baffle step exists and based on prior knowledge and experience I set a target for near field measurement.
That target is what I have drawn for you in earlier posts where bass is 4 dB over treble in level, and I estimate where the transition begins and ends out of which baffle width I have for the project.
So you don't need to compensate for different distances, just hit the target curve with near field measurement.
Once you measure it far field in the room at listening position it should be pretty close to a straight frequency response.
If you have a straight frequency when measuring near field you will lack bass since it will be 3-4 dB to low compared to treble.
It's pretty straight forward and far field measurement is almost impossible to fully trust since reflections and standing wave nodes in the room normally have a huge impact and create lots of peaks and dips in the bass region below 200 Hz.
It took a while to find "my" way of performing measurement since it is an art that is not that easy as a DIY without controlled envioronment like silent room.
I just let you know my experience, but I expect you to find your own way how to perform your measurements.
Hopefully you have got better insight into the problem and can make better choices once you start building your new loudspeaker.
Remember, it's a hobby...
We learn from trial and errors too.
Now I understand your point of view.
In short you are saying to also consider this when making measurements and to do that the only way is to measure at close range.
How to take action is quite simple, but it is worthwhile to do several trials and find the right compromise.
I will certainly consider this aspect as well when I design the next loadspeakers, but I have to find the right way to make the measurements because as you said and as I myself have verified the environmental factors in a room that is not properly treated could falsify the measurements.
In fact, this is one of the aspects that I would like to treat as a priority when I change the house.
I think perhaps the acoustic treatment of a room is more important than the speaker itself.
I would say that measurements in near field between 5-30 cm is the only measurements that will be accurate enough to not be influenced by baffle step and contrubution from the room, so that is what I use for measuring basically everything. Gating helps a bit but is not the perfect solution.
Of course I make hundreds of other measurement on off axis and at different distances to verify and see final result at far field and listening position.
But for critical leveling of drivers, making crossovers etc etc it is near field measurement only.
I measure drivers both individually and then together at 50-90 cm distance to verify that they play well together and behave as expected.
So a target curve that I know will get me in the right ballpark is essential for the process.
Find the target, and then fine tune, listen and verify with near-, mid- & far field measurements if anything needs to be adjusted.
It might be that you end up with too much bass if you use my method since basically bass is elevated 4 dB.
If you measured for a straight frequency in near field with your former design you probably ended up with a bass that actually was 4 dB down
when listened to from your listening position and that you needed that extra 3-4 dB you got from the bass reflex port.
If you elevate with 4 dB you might end with a SPH250 KE in a sealed cabinet for best overall balance. I think that will be a better solution in total, but that is for you to decide during the design process.
Oh yes, acoustic treatment is mandatory, but if you get the speaker fairly right from the beginning it usually is enough to find a good sounding position witheout the need for basstraps etc, etc.
Controlling reflections and reverb time is essential.
Of course I make hundreds of other measurement on off axis and at different distances to verify and see final result at far field and listening position.
But for critical leveling of drivers, making crossovers etc etc it is near field measurement only.
I measure drivers both individually and then together at 50-90 cm distance to verify that they play well together and behave as expected.
So a target curve that I know will get me in the right ballpark is essential for the process.
Find the target, and then fine tune, listen and verify with near-, mid- & far field measurements if anything needs to be adjusted.
It might be that you end up with too much bass if you use my method since basically bass is elevated 4 dB.
If you measured for a straight frequency in near field with your former design you probably ended up with a bass that actually was 4 dB down
when listened to from your listening position and that you needed that extra 3-4 dB you got from the bass reflex port.
If you elevate with 4 dB you might end with a SPH250 KE in a sealed cabinet for best overall balance. I think that will be a better solution in total, but that is for you to decide during the design process.
Oh yes, acoustic treatment is mandatory, but if you get the speaker fairly right from the beginning it usually is enough to find a good sounding position witheout the need for basstraps etc, etc.
Controlling reflections and reverb time is essential.
Last edited:
In my current project at the time I performed various tests and also closed the bass reflex hole.
The biggest differences that occurred with closed or open box were in the range of 30-150 Hz not above.
In the closed box it dropped in frequency lower but lacked that resonance peak around 70/80 Hz that made the bass more powerful.
In a closed box with the bass reflex I sacrificed a few less Hz to have more power around 60/80 Hz.
In short, what happens after 150 Hz is not dependent on the bass reflex, but to other factors such as environmental factors, amount of the sound absorbing material, and apparently also on the baffle step.
If I wanted a flat response on 30/150 Hz then I would have opted for the closed box, but that was not my goal.
I like flat response after 120 Hz not before.
The biggest differences that occurred with closed or open box were in the range of 30-150 Hz not above.
In the closed box it dropped in frequency lower but lacked that resonance peak around 70/80 Hz that made the bass more powerful.
In a closed box with the bass reflex I sacrificed a few less Hz to have more power around 60/80 Hz.
In short, what happens after 150 Hz is not dependent on the bass reflex, but to other factors such as environmental factors, amount of the sound absorbing material, and apparently also on the baffle step.
If I wanted a flat response on 30/150 Hz then I would have opted for the closed box, but that was not my goal.
I like flat response after 120 Hz not before.
Last edited:
Possibly not. The answer is always in the FR/phase plot. Impedance compensation circuits aren't mandatory, but rather very useful in some cases when they make designing the filters easier and behave ideally. I made a number of example plots about it here, I hope it helps you.
With the ready availability of Virtuix and XSim, it's super easy to make this judgement. Sometimes, just sometimes, you can avoid the compensation by just picking different part values. 🙂
First, sorry for -3 vs -6dB discussion we had.
You are perfectly correct about -6dB for each doubling of distance for point source, and on second thought, I know this too.
The measurements I showed from the B&O Penta reduced by -3dB for each doubling of distance and that is of course because it is a linear array as you mentioned. It uses four 5" bass drivers if memory serves.
Now some new measurements as promised:
First picture show difference between SPL and Gated SPL measurements. Basically equal up to 50 cm from baffle.
Please note that very close measurements closer than the drivers diameter are corrupted in the upper regions.
I measure from at least the same distance as the diameter of the driver. So no super close near field measurements.
Second picture shows amplitude level in relation to distance and as you said -6dB per doubling of distance.
In third picture I have scaled all measurements to the same level, and here the effect of the baffle step can be seen where
sound pressure drops faster when the loudspeaker goes from half space to full space distribution. Needs to be compensated for a linear frequency response at listening position. Here the width of baffle are 28cm and edges are sharp 90°
Last edited:
This graph seems not to confirm what has been seen in previous measurements.
In the previews graphs the baffle step issue was more present at closer distances within 50cm and at 110cm, the response was almost flat.
But in this one Instead, it seems to be the opposite, less baffle step issue at 20/40cm and more at 80cm.
Do you have measure the some loudspeakers ? and you have putted the microphone in the some point ?
In the previews graphs the baffle step issue was more present at closer distances within 50cm and at 110cm, the response was almost flat.
But in this one Instead, it seems to be the opposite, less baffle step issue at 20/40cm and more at 80cm.
Do you have measure the some loudspeakers ? and you have putted the microphone in the some point ?